JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1990, p. 2621-26260095-1137/90/122621-06$02.00/0Copyright © 1990, American Society for Microbiology

Vol. 28, No. 12

Coaggregation of Oral Candida Isolates with Bacteria fromBone Marrow Transplant Recipients

LAN YIN HSU,' GLENN E. MINAH,2* DOUGLAS E. PETERSON,3 JOHN R. WINGARD,4 WILLIAM G. MERZ,4VIKI ALTOMONTE,4 AND CAROLYN A. TYLENDA5

Department of Microbiology, National Taiwan University Medical School, Taipei, Taiwan'; Department of Microbiology,Dental School, University ofMaryland at Baltimore, Baltimore, Maryland 212012; Department of Oral Diagnosis, SchoolofDental Medicine, University of Connecticut Health Center, Farmington, Connecticut 060323; Johns Hopkins University

Hospital, Baltimore, Maryland 212054; and Office ofDevice Evaluation, U. S. Food and Drug Administration,Rockville, Maryland 208505

Received 18 June 1990/Accepted 29 August 1990

In vitro coaggregation between Candida species isolated from immunosuppressed bone marrow transplantrecipients and oral bacteria was investigated. Each Candida strain showed a different pattern of coaggregationwith the 22 bacterial strains studied. Two strains of LactobaciUlus amylovorus isolated from separate bonemarrow transplant patients and Fusobacterium nucleatum (VPI 10197) coaggregated with ail Candida strains.Ten bacterial strains showed no coaggregation with the Candida strains. A variety of inhibition patterns wereobserved when coaggregating strains were first incubated with various sugars or subjected to heat treatment.Positive and negative results were generally consistent with ail Candida strains. On the basis of the culturecharacteristics of the oral rinse specimens, relationships between the colonization of bacteria and yeasts and invitro coaggregation were suggested.

Bone marrow transplant (BMT) recipients frequentlyexperience the proliferation of Candida populations in theoropharynx during the period of neutropenia following my-elosuppression (21, 28, 33). Consequences include localizedcandidiasis and, in some patients, fungemia (20, 21, 33).Although the oral environment is profoundly altered by

medications and antimicrobial agents received by patientsduring transplantation, it is not known whether candidalsuccession is primarily a response to these covariates or toother factors, such as a reduction in host immunity, changesin candidal adherence to oral epithelium (18, 19), or bacte-rium-yeast interactions. The last may involve coaggregation(2), growth stimulation, or inhibition (16).The aim of the present investigation was to characterize

oral fungal and bacterial inhabitants of BMT patients and toexamine coaggregation between Candida isolates and oralbacteria. The bacterial strains, chosen to represent commonoral inhabitants of immunosuppressed cancer patients (23,27) and normal subjects (22, 24), were both reference strainsand direct isolates from BMT patients.

MATERIALS AND METHODSSubjects and specimen collection. Eight post-BMT patients

at the Johns Hopkins Oncology Center (JHOC) participated.All were diagnosed as having a hematologic malignancy(acute lymphoblastic leukemia, acute myelogenous leuke-mia, or lymphoma) before receipt of the auto- or allogeneicBMT, and all were neutropenic (<1,000 neutrophils permm3) at the time of specimen collection. Subjects rinsedwith 10 ml of sterilized 0.85% saline solution for 30 s andexpectorated into sterile plastic screw-cap containers. Thecontainers were transferred within 1 h to an anaerobicchamber (Coy Lab Products, Inc., Ann Arbor, Mich.) con-taining a gaseous atmosphere of 5% carbon dioxide-10%hydrogen-85% nitrogen and with an ambient temperature of

* Corresponding author.

37°C. The anaerobe laboratory was located at the Universityof Maryland Dental School.

Culture procedures. The objectives of the culture proce-dures were to (i) describe semiquantitatively the total sali-vary microbiota in the neutropenic subjects, (ii) isolate andcharacterize salivary bacteria appearing in abnormally highconcentrations, and (iii) obtain fungal isolates for the coag-gregation experiments. Definition of abnormally high con-centrations was based on quantitative studies of salivarymicroflora from normal subjects conducted previously in ourlaboratory with similar culture techniques (22-24).Specimens were dispersed for 10 s by ultrasound (Cell

Disruptor; Kontes, Vineland, N.J.), serially diluted in re-duced transport fluid (17), and plated on nonselective andselective agar media. MM10 sucrose blood agar (17) wasused as the nonselective medium. This medium was incu-bated anaerobically and aerobically in 10% carbon dioxide at37°C. Selective media were (i) mitis salivarius-bacitracin-sucrose agar for Streptococcus mutans (10), (ài) actinomycesagar (35), (iii) neisseria agar, (iv) Rogosa SL agar forlactobacilli (Difco, Detroit, Mich.), (v) veillonella agar (Dif-co), (vi) Sabouraud dextrose agar for yeasts (BBL, Cock-eysville, Md.), (vii) mannitol-salt agar for staphylococci(BBL), (viii) desoxycholate agar for enteric bacilli (Difco),(ix) laked sheep blood agar with vancomycin and kanamycinfor oral Bacteroides spp. (30), and (x) crystal violet-erythro-mycin agar for Fusobacterium spp. (34).

Representative fungal isolates from Sabouraud dextroseagar and isolates of bacteria determined to be present inabnormally high concentrations were purified on nonselec-tive medium and identified with appropriate API (AnalytabProducts, Plainview, N.Y.) rapid identification systems.

Coaggregation assay. The coaggregation assay of Bagg andSilverwood (2) was used. Aliquots (0.3 ml) of approximately107 yeast cells per ml (a McFarland turbidity of 4) and 108bacterial cells per ml (a McFarland turbidity of 1) in potas-sium phosphate buffer (0.25 M, pH 7.4) were combined insterile 7-ml screw-cap glass vials and incubated for 12 h at

2621

2622 HSU ET AL.

TABLE 1. Predominant bacteria and yeasts in salivary rinse cultures of BMT patients

Organism (% total viable county)Patient

Predominant bacteria Yeasts

JH01 P. stutzeri (55)b, L. casei subsp. pseudoplantarum (18)" C. albicans (0.03)JH02 GNB (78) C. lambica (0.001)bJH03 GNB (92) NDCJH04 GNB (60), alpha-hemolytic streptococci (40) NDJH05 S. mitis (67)b, P. maltophilia (16)b, L. amylovorus (16.0)b C. albicans (0.7)bJH06 Alpha-hemolytic streptococci (48), L. amylovorus (8.0)b, Veillonella species (36) C. albicans (1.0)bJH07 GNB (80), S. sanguis (20) NDJH08 P. stutzeri (99)b C. albicans (0.01)ba Determined on MM10 sucrose blood agar incubated anaerobically at 37°C.b Used in coaggregation experiments.C ND, Not detected.

room temperature on an incubator-shaker (New BrunswickScientific, Inc., Edison, N.J.) oscillating at 33 rpm. Afterincubation, 1 drop of the mixture was placed on glassmicroscope slides, covered with no. 2 coverslips, and exam-ined by dark-field microscopy at x400. Coaggregation wasobserved as distinct aggregations of yeast and bacterialparticles. Negative control suspensions of either yeasts orbacteria in potassium phosphate buffer were made andexamined for coaggregation as described above. Each test orcontrol reaction was conducted in triplicate.The following yeast strains were used: isolates from the

BMT patient salivary specimens, Candida albicans ATCC18804, a blood isolate of C. albicans from an immunosup-pressed patient at JHOC, a blood isolate of C. tropicalisfrom an immunosuppressed patient at JHOC, and a salivaryC. albicans isolate from an immunosuppressed patient at theUniversity of Maryland Cancer Center.The following reference bacterial strains were used along

with salivary isolates from BMT patients (listed in Results):Lactobacillus salivarius ATCC 11741, L. acidophilus ATCC521, L. casei ATCC 11578, Staphylococcus epidermidisATCC 12228, S. salivarius ATCC 25975, S. sanguis ATCC10556, S. sanguis ATCC 10557, S. sanguis ATCC 903, S.mutans Ingbritt ATCC 1600, Klebsiella pneumonia ATCC13833, Escherichia coli ATCC 25922, Pseudomonas aerugi-nosa ATCC 27853, Actinomyces viscosus ATCC 15987,Bacteroides gingivalis CS213 isolated from a periodontalpocket of a patient at the University of Maryland DentalSchool, and Fusobacterium nucleatum VPI 10197. Thereference bacterial strains were selected to represent indig-enous oral bacteria and species known to inhabit the mouthsof immunosuppressed patients.

Inhibition of coaggregation. The objective of these exper-iments was to identify substances or conditions which inhib-ited coaggregation reactions. Although the present investi-gation was not designed to characterize the biochemistry ofcoaggregation reactions, the inhibition studies were intendedto help categorize and partially elucidate the bacterium yeastinteractions.The following substances or conditions were tested for

their inhibitory effect: (i) D-(+)-glucosamine hydrochloride(Sigma, St. Louis, Mo.), (ii) L-fucose (Sigma), (iii) ot-methyl-mannoside (Sigma), (iv) dextrose (Fisher, Pittsburgh, Pa.),(v) D-mannose (BBL), (vi) ot-methyl-glucoside (Mann Re-search Laboratories, New York, N.Y.), and (vii) heat treat-ment (85°C for 30 min).An inhibition assay described by Critchley and Douglas (5)

was used. The reagents were prepared in 0.25 M potassiumphosphate buffer at a concentration of 200 mg/ml. Samples(50 ,u) of the individual stock solutions were added to 0.2-ml

bacterial suspensions (108 cells per ml) before the addition of0.2-ml yeast suspensions (107 cells per ml). Heat treatmentswere conducted by exposing individual strains to heat beforemixtures were prepared. The culture conditions were thesame as those described above for the coaggregation assay.

RESULTS

Culture findings. The distinctive features of the salivarymicroflora of the BMT patients designated JH01 to JH08 arepresented in Table 1.Gram-negative bacilli (GNB) were predominant in six of

the eight specimens and were identified as Pseudomonasspecies in three specimens. Yeasts either were not detectedin specimens in which GNB were predominant or werepresent in low concentrations. Relatively high levels ofyeasts were recovered in specimens from JH05 and JH06, inwhom GNB either were not detected (JH06) or were notpredominant (JH05). In both of the latter cultures, L. amy-lovorus was isolated.The microbial specimens obtained from the BMT patients

and used in coaggregation experiments were four strains ofC. albicans and one of C. lambica, L. casei subsp. pseudo-plantarum, Pseudomonas stutzeri from two individuals, P.maltophilia, L. amylovorus from two individuals, and S.mitis. Additional candidal and bacterial strains are listed inMaterials and Methods.

Coaggregation assay. The results of the coaggregationassays are shown in Table 2 and Fig. 1. None of the Candidastrains autoagglutinated. The three strains of S. sanguiswere the only bacteria which autoagglutinated. Three strainsof bacteria coaggregated with all of the yeast strains. Thesewere L. amylovorus isolated from two BMT patients and F.nucleatum ATCC 10197. Another Lactobacillus species, L.casei subsp. pseudoplantarum, isolated from a BMT patientshowed no coaggregation with Candida strains, nor didreference strain L. casei ATCC 11578 or L. acidophilusATCC 512. L. salivarius ATCC 11741 coaggregated with twoof the nine yeast strains. S. mitis isolated from a BMTpatient coaggregated with five yeast strains, including ayeast strain from the same patient. S. epidermidis ATCC12228 coaggregated with two yeast strains. S. salivariusATCC 25975 only coaggregated with C. albicans and C.tropicalis isolated from a blood specimen of a BMT patient.S. sanguis ATCC 10556 coaggregated with six yeast strains,and S. sanguis ATCC 10557 coaggregated with seven. AI-though some autoagglutination was seen with S. sanguis,positive and negative interactions with yeast strains could bedetected by examining the cellular morphology of thestrains. A. viscosus ATCC 15987 and B. gingivalis CS213

J. CLIN. MICROBIOL.

COAGGREGATION OF ORAL CANDIDA ISOLATES WITH BACTERIA

TABLE 2. Coaggregation of Candida species with oral bacteria

Coaggregation of:Autoag-

Oral bactenum C. aibi- C. lam- C. aibi- C. aibi- C. aibi- C. aibi- C. aibi- C. tropi- C. aibi- glutina-cans bica cans cans canis canis bascls cn ino(JHO1)(JHO2)(UMCC) (JHO5) (JHO6) (JHO8) (JH, (JH, (ATCC bacteriab(JH01)(JH02)(UMCQ JH05) JH06) JH08) blood) blood) 18804)

L. casei subsp. pseudo-plan- ----- - ----tarum (JH01)

Pstutzeri(JHO1)u-PL. amylovorus (JH+5) + + + + + + + + +-S. mitis (JH05) + + + + - - - - + +PO.S)a-t-o-h--i-J-H- --L.aamylovorus (JH+6) + + + + + + + + +-P. stutzeri (JH08) - - - - - - - - - -L. salivarius (ATCC 11741) - - - - - - + - + -L.aacidophilus (ATCC-521) - - - - - -

L.ccasei(ATCC 11578 - - - - - -

S. epidermidis (ATCC 12228) - - + - - + - + - -+S.5salivarjus (ATCC 25975)u-SS. sanguis (ATCC 10556) - - - + + + + + + +S. sanguis (ATCC 10557) + - + + + + + - + +S. sanguis (ATCC 903) + - + - - + + - + +S. mutans Ingbritt (ATCC 1600) - - - - - - - - - -

K. pneumoniae (ATCC 13883) - - - - - - - - - -

E. col-(ATCC 25922) - - - - - - - - - -

P. aeruginosa (ATCC 27853) - - - - - - - - - -

A. viscosus (ATCC 15987) + - - + + - + + - -

B. g es-CS21-)+- + + + - - - - - -F. nucleatum (VPI 10197) + + + + + + + + + +

a Designations in parentheses indicate strain designation, BMT patient, orb There was no autoagglutination of yeasts.

coaggregated with five and three yeast strains, respectively.L. casei subsp. pseudoplantarum isolated from a BMTpatient, P. maltophilia, and P. stutzeri isolated from sepa-rate BMT patients, L. acidophilus ATCC 521, L. caseiATCC 11578, S. mutans Ingbritt ATCC 1600, K. pneumo-niae ATCC 13883, Escherichia coli ATCC 25922, and P.aeruginosa ATCC 27853 did not coaggregate with any yeaststrains.

Inhibition of coaggregation. Yeast-bacterium mixtures

FIG. 1. Coaggregation between C. albicans and A. viscosus.

other source. UMCC, University of Maryland Cancer Center.

which exhibited coaggregation were incubated with a varietyof sugars and subjected to heat prior to mixing (85°C for 30min) in an effort to examine the specificities of the coaggre-gation reactions (Table 3).

All Candida strains which coaggregated with a particu-lar bacterial strain reacted similarly in the inhibition exper-iments when combined with the same bacterial strain. Re-actions involving all S. sanguis strains, S. epidermidis, andA. viscosus were not inhibited by the sugars tested. Reac-tions involving S. mitis, L. salivarius, S. salivarius, and B.gingivalis were inhibited by only one sugar each. Thesewere, in respective order, a-methyl-glucoside, a-methyl-glucoside, D-(+)-glucosamine, and dextrose. Coaggregationreactions between Candida strains and F. nucleatum wereinhibited by a-methyl-mannoside and D-mannose. Thosebetween Candida strains and both L. amylovorus strainswere inhibited by oa-methyl-mannoside, D-mannose, dex-trose, and a-methyl-glucoside.

After heat treatment, S. salivarius, S. epidermidis, B.gingivalis, and F. nucleatum stiil coaggregated with yeaststrains. However, this activity was lost in reactions involv-ing L. amylovorus, S. mitis, L. salivarius, all S. sanguisstrains, and A. viscosus.

DISCUSSIONThe experimental results confirmed the report of Bagg and

Silverwood (2) that certain strains of indigenous oral bacteriabound to C. albicans and formed visible aggregates. In theirstudy, strains of S. sanguis NCTC 10558, S. salivariusNCTC 8606, S. mutans D282 (NCTC 10832), S. mitis 3(NCTC 10712), F. nucleatum NGB 15-73, and A. viscosusATCC 15987 coaggregated with C. albicans NCPF 3281 (2).The results of the present study are consistent with thefindings of Bagg and Silverwood (2), with the exception that

VOL. 28, 1990 2623

2624 HSU ET AL.

TABLE 3. EffeCt of sugars and heat on the coaggregation of Candida species with oral bacteria

Coaggregation of all Candida isolates after treatment with:Oral

bacterium Nothing a-Methyl- D- L- D-(+)- Dextrose -Methyl- Heat(control) mannoside Mannose Fucose Glucosamine glucoside

L. amylovorus (JHO5) + - - + +S. mitis (JHO5) + + + + + + - -L. amylovorus (JHO6) + - - + +L. salivarius (ATCC 11741) + + + + + +S. salivarius (ATCC 25975) + + + + - + + +S. epidermidis (ATCC 12228) + + + + + + + +S. sanguis (ATCC 10556) + + + + + + + -

S. sanguis (ATCC 10557) + + + + + + + -

S. sanguis (ATCC 903) + + + + + + + -

A. viscosus (ATCC 15987) + + + + + + + -B. gingivalis (CS213) + + + + + - + +F. nucleatum (VPI 10197) + - - + + + + +

a Designations in parentheses indicate strain designation or BMT patient.

S. mutans Ingbritt and E. coli ATCC 25922 did not coaggre-gate with any Candida strain. As wide differences in coag-

gregation occurred with Candida strains in the presentstudy, selected bacterial strains may also react differently.The E. coli strain used by Bagg and Silverwood (2) was a

fimbriated strain which coaggregated consistently with theirCandida strain. The type of fimbriae present on the E. colistrain used in the present study or their absence may explainour negative findings. Other possible explanations for differ-ences in coaggregation between various investigations aredifferences in the surface properties of Candida strainsinduced by various growth media (4, 19) or growth condi-tions (4, 14). A recent study by Brawner and Cutler (3) foundthat C. albicans strains isolated from immunosuppressedindividuals differed serologically from strains isolated fromimmunocompetent individuals.Attachment of microorganisms to epithelial cells is recog-

nized as a crucial step in the colonization and infection of thehost. Although it was suggested that certain bacteria possessligands that could bind to both candidal and epithelial cellsand thereby mediate adherence of the yeasts by a "bridging"action, the possible role of this type of attachment in relationto oral colonization by yeasts has received little study.Gibbons and van Houte (9) reported that S. salivarius and

S. sanguis were recovered in high proportions from cheekand tongue surfaces, which are often colonized by C. albi-cans, whereas S. mutans was present only in low propor-tions, if at all, on these surfaces. The coaggregation found inthe present study between candidal strains and S. sanguisand S. salivarius but not S. mutans supports the possibilitythat certain streptococci may act as mediators of candidalattachment to oral mucosal surfaces.

Previous investigations have not examined the coaggrega-tion of Lactobacillus species with yeast cells. We observedthat two strains of L. amylovorus isolated from BMT pa-tients coaggregated with all nine yeast strains. It is possiblethat L. amylovorus, previously reported to be isolated fromwaste corn fermentation (29), plays an important role incandidal colonization. Our results showing that L. salivariusATCC 11741 coaggregated with yeast strains while L. aci-dophilus ATCC 521, L. casei ATCC 11578, and L. caseisubsp. pseudoplantarum did not suggest that L. salivariusmay aid candidal cells in mucosal colonization of the oralcavity surfaces which are in contact with saliva, while L.salivarius is commonly isolated from the saliva and tongue.L. acidophilus is usually found in dental plaque, and L. caseiis usually found in carious dentin (11).

A. viscosus, a predominant Actinomyces species isolatedfrom dental plaque, is recognized as being indigenous to thehuman oral cavity (12). A. viscosus ATCC 15987 coaggre-gated with C. lambica, C. tropicalis, and two strains of C.albicans in the present study. Bagg and Silverwood (2) alsoreported coaggregation between Actinomyces and Candidastrains.

F. nucleatum VPI 10197, isolated from the gingival crev-ice, has shown a characteristic hemagglutination patternwith a variety of erythrocytes of human and animal origin(8). It was suggested that this bacterial species possesseslectins that recognize galactoselike residues on human cellsurfaces and possibly allow attachment of the bacteria to cellsurfaces. Since coaggregation was found in the present studybetween F. nucleatum VPI 10197 and all the yeast strainstested, a potential for mediation by this bacterium of oralcolonization by yeasts exists.

Bacteriodes species have been recognized as prominentmembers of maturing dental plaque (31). B. gingivalis wasreported to be the only member of the black-pigmentedBacteroides group that coaggregated with several oral spe-cies, including Veillonella, Capnocytophaga, Actinomyces,and Rothia species (6). In the present study, B. gingivalisCS213 coaggregated with three yeast strains. B. gingivaliswas reported to be present on the dorsum of the tongue andoral mucosa but only in very small numbers (15). Bagg andSilverwood (2) proposed that, because of the scarcity of B.gingivalis, its cooperation in the colonization of the oralmucosa by yeast strains in vivo is unlikely.

Previous investigations have suggested that differences inenzyme production or the ability to adhere to epithelial cellsexist among various strains of C. albicans (1). Our findingsof differences in coaggregation among seven strains of C.albicans strongly suggest that specificities of coaggregatingpairs existed at the strain level. Differences also existed atthe species level. C. lambica, for example, showed lesscoaggregation with bacterial strains than did C. albicans orC. tropicalis.Organisms with the ability to coaggregate with epithelial

surfaces or plaque bacteria may have an advantage overnoncoaggregating organisms, which could be more easilyremoved from the oral environment by salivary flow. Baggand Silverwood (2), however, proposed that large aggregatesof bacteria and yeasts may be more readily cleared from theoral cavity than are individual cells. In vivo studies involvinggerm-free animals may be necessary to elucidate the biolog-ical significance of in vitro coaggregation.

J. CLIN. MICROBIOL.

COAGGREGATION OF ORAL CANDIDA ISOLATES WITH BACTERIA

There appeared to be consistent correlations betweensome of the culture findings and coaggregation reactions.Specimens which yielded GNB, usually Pseudomonas spe-cies, as predominant isolates harbored only low levels ofCandida organisms or no Candida organisms (JH01, 0.03%;JH02, 0.01%; JH03, 0.0%; JH04, 0.0%; JH07, 0.06%; andJH08, 0.01%). As none of the Candida strains coaggregatedwith the Pseudomonas strains from the BMT patients orreference strains, perhaps oral mucosal sites occupied bycertain Pseudomonas species are not accessible to coloniza-tion by Candida species. Another possibility, that Pseudo-monas species actively inhibit Candida species, should beinvestigated.Specimens showing relatively high levels of Candida

organisms contained bacterial strains with which the samecandidal strains coaggregated in vitro. JH05, with 0.7% C.albicans, harbored a high concentration of L. amylovorus.JH06, with 1.0% C. albicans, harbored L. amylovorus andvery high concentrations of a catalase-positive Actinomycesspecies. If such bacterial species are attached to oral muco-sal sites, adherence to those sites by Candida species may beenhanced.An attempt was made to determine the mechanism of

coaggregation of candidal cells with bacterial cells by inhi-bition with sugars or heat treatment. Previous studies havefound that the outermost cell wall of C. albicans containsmannoselike receptors which are probably involved in yeastattachment to buccal cells (13). Mirelman et al. (25) alsodemonstrated that many bacteria have the ability to bind tomannose-containing receptors such as yeast mannans, whichare the major polymers of the yeast cell wall. Strains of L.amylovorus, S. mitis, and F. nucleatum were inhibited fromcoaggregation with yeast strains by mannose in the presentstudy. This result suggests that the above-listed bacteriahave mannose-binding properties. S. sanguis, A. viscosus,S. salivarius, B. gingivalis, and L. salivarius were consid-ered negative with respect to mannose-binding capacity.a-Methyl-mannoside, a polymer ofa-methyl-mannose whichdiffers from D-mannose in that the OH group at carbon 1 inD-mannose is replaced by a CH3 group, displayed the sameinhibitory effect as D-mannose.

D-(+)-Glucosamine, a component of chitin, did not showan inhibitory effect, except with S. salivarius. This resultwas expected, since chitin is thought to be concentrated inthe inner layers of the yeast cell wall and thus to beunavailable for surface interactions (32). The exception of S.salivarius cannot be explained.

L-Fucose, which has been shown to have no inhibitoryeffect on F. nucleatum hemagglutination activity (26), dis-played no inhibition of candidal coaggregation with F. nu-cleatum or other bacteria tested in this study. This resultsuggests the absence of L-fucose-like groups at the termini ofattachment sites.Dextrose and a-methyl-glucoside were equally capable of

inhibiting L. amylovorus coaggregation with yeast cells.Neither of them inhibited yeast coaggregation with S. epi-dermidis, S. salivarius, S. sanguis, or F. nucleatum. Thesetwo sugars had different effects on the coaggregation activityof S. mitis, B. gingivalis, and L. salivarius in that dextroseinhibited B. gingivalis and a-methyl-glucoside inhibited S.mitis and L. salivarius from coaggregation with yeast cells.Eshdat et al. (7) isolated a mannose-binding lectin (carbo-

hydrate-binding protein) from E. coli. This bacterial proteinwas proposed to play a role in the agglutination of yeastcells by E. coli, a process which could be reversed byheat treatment. As our results showed that heat treatment

abolished the coaggregation activities of L. amylovorus,L. salivarius, S. sanguis, S. mitis, and A. viscosus, it issuggested that a certain bacterial or yeast protein(s) partic-ipated in the coaggregation mechanism. A role of bacterial oryeast protein in mediating yeast coaggregation is evidentonly when heat treatment inhibits coaggregation. This wasthe case with S. sanguis and A. viscosus. A recent study byJenkinson et al. supported the presence of a protein mediatorin coaggregation reactions between Candida species andcertain Streptococcus species (14).

Previous investigations showed that bacterial adhesion tomammalian cells occurs by various means, including lectin-like interactions, cell surface adhesins, hydrophobicity, andfibronectin (18). Although the tests in the present study werenot performed to determine the exact mechanism(s) ofcoaggregation, the results indicated that lectinlike interac-tions and nonmannoprotein adhesins contribute to the coag-gregation of yeast cells with bacterial cells.As previous coaggregation studies did not address the

quantities of yeasts and bacteria in natural specimens, theresults cannot be compared with those of other research. Itis recognized that the BMT subjects presented many com-plex variables, such as antifungal, antibiotic, and immuno-suppressive treatment regimens, as well as different degreesof immunosuppression. These, of course, make the estab-lishment of definitive conclusions difficult. The data do,however, indicate that oral bacteria should be regarded asfactors which may contribute to Candida colonization andproliferation in the oral cavities of BMT subjects.

ACKNOWLEDGMENT

This research was supported by the University of Maryland atBaltimore.

LITERATURE CITED1. Allen, C. M., and F. M. Beck. 1987. Differences in mucosal

reactions related to Candida albicans isolates. J. Oral Pathol.16:89-93.

2. Bagg, J., and R. W. Silverwood. 1986. Coagglutination reactionsbetween Candida albicans and oral bacteria. J. Med. Microbiol.22:165-169.

3. Brawner, D. L., and J. E. Cutler. 1989. Oral Candida albicansisolates from nonhospitalized normal carriers, immunocompe-tent hospitalized patients, and immunocompromised patientswith and without acquired immunodeficiency syndrome. J. Clin.Microbiol. 27:1335-1341.

4. Chaffin, W. L., J. Skudlarek, and K. J. Morrow. 1988. Variableexpression of a surface determinant during proliferation ofCandida albicans. Infect. Immun. 56:302-309.

5. Critchley, I. A., and L. J. Douglas. 1987. Role of glycosides asepithelial ceIl receptors for Candida albicans. J. Gen. Micro-biol. 133:637-643.

6. Eke, P. I., V. O. Rotimi, and B. E. Laughon. 1989. Coaggrega-tion of black-pigmented Bacteroides species with other oralbacteria. J. Med. Microbiol. 28:1-4.

7. Eshdat, Y., I. Ofek, Y. Yashouv-Gan, N. Sharon, and D.Mirelman. 1978. Isolation of a mannose-specific lectin fromEscherichia coli and its role in the adherence of the bacteria toepithelial cells. Biochem. Biophys. Res. Commun. 85:1551-1559.

8. Falkler, W. A., Jr., and C. E. Hawley. 1977. Hemagglutinatingactivity of Fusobacterium nucleatum. Infect. Immun. 15:230-238.

9. Gibbons, R. J., and J. van Houte. 1971. Selective bacterialadherence to oral epithelial surfaces and its role as an ecologicaldeterminant. Infect. Immun. 3:567-573.

10. Gold, O. C., H. V. Jordan, and J. van Houte. 1973. A selectivemedium for Streptococcus mutans. Arch. Oral Biol. 18:1357-1365.

VOL. 28, 1990 2625

2626 HSU ET AL.

11. Hoffman, H. 1982. Gram-positive bacterial rods: Clostridium,Bacillus, Lactobacillus, and Listeria, p. 336-338. In W. A.Nolte (ed.), Oral microbiology. C.V. Mosby Co., St. Louis.

12. Hoffman, H. 1982. Actinomyces and related microorganisms, p.426-432. In W. A. Nolte (ed.), Oral microbiology. C.V. MosbyCo., St. Louis.

13. Howlett, J. A., and C. A. Squier. 1980. Candida albicansultrastructure: colonization and invasion of oral epithelium.Infect. Immun. 29:252-260.

14. Jenkinson, H. F., H. C. Lala, and M. G. Shepherd. 1990.Coaggregation of Streptococcus sanguis and other streptococciwith Candida albicans. Infect. Immun. 58:1429-1436.

15. Kelstrup, J. 1966. The incidence of Bacteroides melaninogeni-cus in human gingival sulci and its prevalence in the oral cavityat different ages. J. Periodontol. 4:14-18.

16. Liljemark, W. F., and R. J. Gibbons. 1972. Proportional distri-bution and relative adherence of Streptococcus miteor (mitis) onvarious surfaces in the human oral cavity. Infect. Immun.6:852-859.

17. Loesche, W. J., R. N. Hockett, and S. A. Syed. 1972. Thepredominant cultivable flora of tooth surface plaque removedfrom institutionalized subjects. Arch. Oral Biol. 17:1311-1325.

18. Macura, A. B. 1987. Hydrophobicity of Candida albicans re-lated to their adherence to mucosal epithelial cells. Zentralbl.Bakteriol. Mikrobiol. Hyg. Ser. A 266:491-496.

19. McCourtie, J., and J. L. Douglas. 1984. Relationship betweencell surface composition, adherence, and virulence of Candidaalbicans. Infect. Immun. 45:6-12.

20. Meyers, J. D. 1986. Infection in bone marrow transplant recip-ients. Am. J. Med. 81:27-38.

21. Meyers, J. D., and K. Atkinson. 1983. Infection in bone marrowtransplantation. Clin. Haematol. 12:791411.

22. Minah, G. E., G. B. Lovekin, and L. Chen. 1981. Changes in theoral microflora as indicators of cariogenic potential, p. 55-78. InJ. J. Hefferen, W. A. Ayer, and H. M. Koehler (ed.), Proceed-ings of the Fourth Annual Foods, Nutrition and Dental HealthWorkshop Conference, vol. 3. American Dental Association,Chicago.

23. Minah, G. E., J. L. Rednor, D. E. Peterson, C. D. Overholser,L. G. Depaola, and J. B. Suzuki. 1986. Oral succession ofgram-negative bacilli in myelosuppressed cancer patients. J.Clin. Microbiol. 24:210-213.

24. Minah, G. E., E. S. Solomon, and K. Chu. 1985. The associationbetween dietary sucrose consumption and bacterial populationshifts at six oral sites in man. Arch. Oral Biol. 30:397-401.

25. Mirelman, D., G. Altmann, and Y. Eshdat. 1980. Screening ofbacterial isolates for mannose-specific lectin activity by agglu-tination of yeasts. J. Clin. Microbiol. 11:328-331.

26. Mongiello, J. R., and W. A. Falkler, Jr. 1979. Sugar inhibition oforal Fusobacterium nucleatum haemagglutination and cell bind-ing. Arch. Oral Biol. 24:539-545.

27. Peterson, D. E., G. E. Minah, C. D. Overholser, J. B. Suzuki,L. G. DePaola, D. M. Stansbury, L. J. Williams, and S. C.Schimpff. 1987. Microbiology of acute periodontal infection inmyelosuppressed cancer patients. J. Clin. Oncol. 5:1461-1468.

28. Sanford, G. R., W. G. Merz, J. R. Wingard, P. Charache, and R.Sara. 1980. The value of fungal surveillance cultures as predic-tors of systemic fungal infections. J. Infect. Dis. 142:503-509.

29. Sneath, P. H. A., N. S. Mair, M. E. Sharpe, and J. G. Holt. (ed.).1986. Bergey's manual of systematic bacteriology, vol. 2., p.1219-1234. The Williams & Wilkins Co., Baltimore.

30. Sutter, V. L., D. M. Citron, M. A. C. Edelstein, and S. M.Finegold. 1986. In Wadsworth anaerobic bacteriology manual,4th ed., p. 112-113. Star Publishing Co., Belmont, Calif.

31. Theilade, E., W. H. Wright, S. B. Jensen, and H. Loe. 1966.Experimental gingivitis in man. IH. A longitudinal clinical andbacteriological investigation. J. Periodontal Res. 1:1-13.

32. Tronchin, G., D. Poulain, J. Herbaut, and J. Biguet. 1981.Localization of chitin in the cell wall of Candida albicans bymeans of wheat germ agglutinin. Fluorescence and ultrastruc-ture studies. Eur. J. Cell Biol. 26:121-128.

33. Wade, J. C., and J. C. Schimpff. 1985. Epidemiology andprevention of candida infections, p. 111-133. In G. P. Bodeyand V. Fainstein (ed.), Candidiasis. Raven Press, Publishers,New York.

34. Walker, C. B., D. Ratliff, D. Muller, R. Mandell, and S. S.Socransky. 1979. Medium for selective isolation of Fusobacte-rium nucleatum from human periodontal pockets. J. Clin.Microbiol. 10:844 849.

35. Zylber, L. J., and H. V. Jordan. 1982. Development of aselective medium for detection and enumeration ofActinomycesviscosus and Actinomyces naeslundii in dental plaque. J. Clin.Microbiol. 15:253-259.

J. CLIN. MICROBIOL.